Starburst Data

1.1 Identity and business model

Starburst Data, Inc. is described by two independent company profiles as a private Boston software company founded in 2017. Its core proposition is not another system of record: it commercializes the open-source Trino query engine so enterprises can query distributed data across clouds, data centers, warehouses, lakes, and operational databases. Starburst Galaxy is the managed-cloud offer, while Starburst Enterprise is self-hosted and aimed at private-cloud, hybrid, on-premises, and air-gapped environments. The official product page claims more than 50 native connectors and emphasizes querying data without moving or copying it. That architecture is well aligned with regulated and heterogeneous estates, but paid differentiation still depends on governance, support, performance, and AI features being valuable beyond freely available Trino.[CO001, CO002, CO003, CO004]

1.2 Leadership and governance

Justin Borgman is identified as co-founder and CEO, with prior operating experience at Hadapt and Teradata. The supplied leadership profile identifies Steven Chung as president overseeing operations, go-to-market, and customer success; Toni Adams as the senior partner and alliances leader; and Megan Maslanka as chief people officer. This roster provides visible coverage across product vision, commercial execution, partnerships, and organizational development. It does not establish the board roster, investor directors, voting rights, founder ownership, succession planning, or whether the listed team is complete as of the run date. Those omissions matter because the company remains private and the evidence ties its origin and public identity closely to Borgman.[CO005, CO006, CO007, CO016]

1.3 Financing and scale signals

A 2026 secondary profile reports a $250 million Series D in May 2025 at a $3.35 billion valuation, led by Alkeon with participation from Andreessen Horowitz, Coatue, Index Ventures, and Salesforce Ventures. Another profile reports more than $414 million raised in total, but its page contains an internally inconsistent funding table, so that total should not be treated as reconciled. SiliconANGLE separately reported a strategic Citi investment without an amount or terms. Two profiles report that Starburst crossed $100 million in annual recurring revenue in February 2026, alongside approximately 40% year-over-year growth, 130% net dollar retention, and a $20 million AI run rate. These are attributed operating metrics, not audited revenue, cash, or profitability measures.[CO008, CO009, CO010, CO011, CO012, CO013]

1.4 Milestones and diligence boundary

The retained chronology links the 2017 founding to a documented two-product structure by mid-2023, a major financing and AI-product push in May 2025, a commissioned economic validation in January 2026, and reported ARR and workforce scale in early 2026. The evidence also introduces caution. Two profiles disagree on 2026 headcount, reporting 507 and 537 employees, while PeerSpot's October 2026 comparison gives Starburst only two reviews and notes greater deployment complexity than Databricks. The pool supplies no verified current customer count, complete cap table, board register, audited statements, gross margin, cash balance, burn, debt schedule, retention cohorts, or customer concentration. It also does not prove the absence of lawsuits, outages, breaches, sanctions, or other adverse events.[CO012, CO013, CO014, CO015, CO016]

1.5 Exhibits

2.1 Market boundary and economic job

The relevant market is enterprise software and services that query, govern, optimize, or expose data across distributed databases, object stores, warehouses, and lakehouse tables without requiring wholesale centralization. The boundary includes Starburst Enterprise, Starburst Galaxy, connectors, governance, workload management, data products, and related implementation or managed-service spend. It excludes general storage capacity, standalone BI seats, generic ETL unrelated to federated access, and all of a cloud warehouse's consumption spend when that spend is not substituting for distributed access. Trino documentation describes the underlying category as a query engine rather than a database system, while independent comparison evidence frames Starburst as a lakehouse/federation platform rather than a conventional warehouse. This boundary matters because the buyer is paying to reduce data movement, integration friction, governance duplication, or migration risk, not simply to purchase more storage.[CM001, CM002, CM003, CM016]

2.2 Sizing lenses and limits

The strongest quantified lens is the global data-virtualization market: Mordor Intelligence estimates USD 6.25 billion in 2025, USD 7.46 billion in 2026, and USD 18.09 billion by 2031, implying a 19.38% 2026–2031 CAGR. Its segmentation provides useful demand shape—cloud represented 69.85% of 2025 revenue, large enterprises 58.25%, and BFSI 31.12%—but these are slices of the same market, not additive TAM, SAM, or SOM layers. A second data-lakehouse report identifies the relevant deployment and buyer segments but does not expose a usable value in the fetched text. Thus the chapter uses multiple lenses while marking Starburst-specific SAM, SOM, share, and obtainable annual spend as unresolved. The sizing pyramid is a directional evidence map, not a mathematically comparable funnel.[CM004, CM005, CM006, CM007, CM015]

2.3 Buyer, user, and payer map

The buying center is shaped by where data resides and who bears the cost of movement, governance, or operational complexity. Platform and data engineering teams are the primary users and technical evaluators; CIO, CTO, CDO, security, risk, and compliance leaders can own budget when sovereignty or regulated-data requirements dominate. Analysts, BI teams, and increasingly AI application teams are downstream users. Adoption is most credible where an enterprise has multiple clouds or legacy systems, cannot freely move sensitive data, or needs one governed access layer. Starburst's public product surfaces emphasize self-managed deployment for on-premises and hybrid estates, and an independent technical review says those deployments require infrastructure, networking, security, and governance expertise. SiliconANGLE also reports a Citi use case across 165 countries. These signals support an enterprise land motion, but they do not establish contract size, conversion rates, or repeatable payback.[CM008, CM009, CM010, CM011]

2.4 Growth drivers, constraints, and valuation relevance

Demand drivers are AI workloads, real-time analytics in regulated industries, data-mesh or logical-fabric architectures, and pressure to limit duplication and egress. Mordor's report identifies each as a market driver and separately flags governance failures, scarce optimization skills, egress fees, and fragmented sovereignty rules as restraints. Product evidence also indicates a switching-cost trade: federation can avoid migration, but deployment, query tuning, security integration, and resource management still require scarce expertise. Mordor reports that 90% of organizations experience critical talent gaps that add three to ten months to digital projects, and an independent technical review distinguishes Galaxy's lower operating burden from Enterprise's infrastructure, networking, security, and governance requirements. Accordingly, valuation relevance depends on proving durable expansion within regulated and distributed estates while validating implementation burden, gross margin, usage economics, and whether customers consolidate onto a competing warehouse or retain multi-platform architectures.[CM011, CM012, CM013, CM014]

2.5 Exhibits

3.1 Competitive landscape and buyer choice

Starburst sits between an open-source query engine and a full data platform. Its Enterprise product is a supported distribution of Trino that queries data lakes, warehouses, operational databases, and on-premises systems in place, while Galaxy supplies a managed cloud path. The direct commercial comparison is Dremio, which also sells federated lakehouse access, but the broader substitute set includes Snowflake and Databricks when the buyer prefers a centralized managed platform, Firebolt or StarRocks when predictable low-latency analytics matters more than federation, and Trino when an internal team can self-manage open source. Spark is adjacent rather than directly equivalent: the available Trino comparison positions Spark for ETL/ELT and Trino for interactive analytics. This means the relevant buying question is architecture and operating model, not a generic feature count.[CP001, CP002, CP003, CP004, CP005]

3.2 Capability, distribution, and trust comparison

Starburst's material product advantage is the combination of in-place federation, deployment flexibility, and enterprise governance. Official product material describes 50+ native connectors, ANSI SQL access, private-cloud, hybrid, on-premises, and air-gapped deployment, plus row- and column-level controls and audit integrations. Those features are relevant to regulated buyers whose data residency or legacy-system constraints make a warehouse migration costly. They are not equivalent to independently verified performance or adoption: the official page's claim of up to 2x faster performance than open-source Trino is a vendor claim, and the independent review recommends workload-specific benchmarking. Databricks is reported to have an advantage in ease of deployment and customer service, while Starburst is favored for features in complex environments. The resulting distribution moat is strongest where partners, security teams, and platform engineers value compatibility, but weaker where a buyer prioritizes a single managed warehouse or a simpler operating model.[CP006, CP007, CP008, CP009, CP010]

3.3 Switching costs, pricing, and moat durability

Switching costs are asymmetric. Moving from Starburst to open-source Trino is relatively plausible because the commercial product is built on the same engine, SQL dialect, and connector model; the sacrifice is enterprise acceleration, governance, provisioning, and support. Moving to storage-centric systems is more disruptive because data must be ingested into a target storage format rather than queried in place. Starburst's freemium and credit-based entry lowers evaluation friction, but it also makes production spend workload-dependent and therefore difficult to compare with fixed subscription alternatives without a customer benchmark. Independent review evidence flags resource consumption, setup complexity, and performance issues on complex queries as potential drawbacks. The moat is therefore durable around regulated federation and existing Trino expertise, but vulnerable to commoditization of connectors, open-source self-management, and buyer decisions to accept migration in exchange for simpler centralized operations.[CP011, CP012, CP013, CP014]

3.4 Exhibits

The positioning map uses ordinal, evidence-backed coordinates rather than a quantitative market benchmark. Higher federation breadth reflects documented multi-source access; higher managed simplicity reflects the degree to which the cited source describes infrastructure as managed rather than self-operated. Starburst is placed between Trino and centralized managed platforms because the evidence supports both a managed Galaxy path and a self-managed Enterprise path. The map should be read as a decision aid for architecture fit, not as a claim that any vendor is objectively faster, cheaper, or more widely adopted.[CP002, CP006, CP011, CP014]

3.5 Exhibits

4.1 Revenue model and monetization

Starburst sells a commercial layer around Trino through Starburst Galaxy, a managed service, and Starburst Enterprise, a self-managed deployment for private cloud, hybrid, and on-premises environments. The company describes Enterprise as connecting to more than 50 sources through one SQL interface without moving or copying data, while its managed offer shifts infrastructure operations to Starburst. That product split supports two revenue motions: recurring managed consumption and enterprise software/support contracts. Public pricing is not a transparent list price. SelectHub reports a starting point of $1 per credit in prose but $0.50 in its pricing panel, so even the public starting point is internally inconsistent. G2 reports an average implementation time of four months, a 13-month return-on-investment figure, and a 10% average discount; these are marketplace or review-derived buyer indicators, not realized Starburst revenue. The FY25 company release reports ARR per customer above $325,000 and an eight-figure-per-year multi-year contract, but does not disclose contract accounting, recognized revenue, or product mix. This combination suggests enterprise-value selling with potentially meaningful services and support obligations, but the mix and revenue-recognition profile remain unverified.[CI001, CI002, CI003, CI004, CI010, CI011]

4.2 Traction and sales-efficiency proxies

The strongest public traction evidence is company-issued rather than audited financial disclosure. Starburst says FY25 net-new customers grew 20%, Galaxy customer growth rose 76% year over year, Galaxy adoption increased 94% year over year, and ARR per customer exceeded $325,000. It also says ten of the top fifteen global banks are customers and that Dell selected Starburst as the analytics query engine for Dell Data Lakehouse. These disclosures indicate enterprise distribution and a growing cloud motion, but they do not establish cohort retention, gross retention, net retention, CAC, payback, sales-cycle duration, or the share of bookings converted to recognized revenue. Independent review evidence is directionally positive but mixed on economics: G2 users cite integration and query performance, while also describing configuration complexity and pricing that can become harder to justify as usage scales. The appropriate underwriting conclusion is commercial traction with incomplete efficiency visibility, not a proven efficient-growth profile.[CI005, CI006, CI007, CI008, CI009]

4.3 Cost structure and capital adequacy

Starburst's cost structure is likely shaped by distributed-query compute, connector and support engineering, enterprise implementation, and cloud operations, but the assigned evidence does not provide gross margin, hosting cost, support cost, capex, working-capital needs, cash balance, debt, burn, or runway. The platform can reduce customer data movement and duplication, yet that customer benefit does not establish Starburst's own margin profile. G2 reviewers report that consumption-based pricing can make monthly expense hard to predict, while PeerSpot characterizes Enterprise deployment as requiring more technical knowledge than Databricks; both are buyer-side signals that workload and delivery economics require measurement rather than inference. The 2025 Citi strategic investment is described by SiliconANGLE, but its size, valuation, instrument, and use of proceeds are not disclosed in the retained sources. Accordingly, capital adequacy cannot be underwritten from funding chronology alone: the next diligence step is a current cash-and-burn schedule reconciled to contracted ARR, renewal timing, implementation backlog, debt, and any financing covenants.[CI010, CI011, CI012, CI013, CI016]

4.4 Financial underwriting verdict

Starburst has credible enterprise traction and a monetization model that can support large annual contract values, but the public record is materially stronger on bookings-adjacent indicators than on accounting quality or cash durability. The reported ARR-per-customer figure and record contract are company claims, and the available independent sources mostly describe product experience or list-price proxies rather than audited financial outcomes. The adverse evidence also matters: an independent sovereignty analysis argues that Galaxy's US control-plane exposure can force regulated buyers toward self-hosting, potentially increasing sales friction, deployment cost, or deal qualification requirements in sensitive sectors. That does not prove lost revenue, but it is a plausible margin and cycle risk. A defensible financial stance is research-more: validate ARR reconciliation, renewal and concentration data, realized discounts, gross margin by deployment type, cash runway, debt, and the economics of the Galaxy-to-Enterprise mix before assigning a stronger recommendation.[CI014, CI015, CI016]

4.5 Exhibits

5.1 Product surface and customer workflow

Starburst's core customer workflow is to connect existing lakes, warehouses, operational databases, and on-premises systems, then query them through a common SQL layer rather than first centralizing every dataset. The open-source foundation is Trino, whose own site describes both in-place analysis and federation across multiple systems. Starburst Enterprise adds a self-managed deployment surface for private-cloud, hybrid, on-premises, and air-gapped environments, while older secondary documentation describes Galaxy as the managed SaaS counterpart. The commercial value therefore sits above the base engine: connector packaging, governance, acceleration, support, and reduced operating burden. A commissioned 2026 validation reports materially shorter source onboarding for interviewed customers, but that evidence should be treated as sponsored customer testimony, not as a representative implementation benchmark.[CE001, CE002, CE003, CE005, CE007, CE008]

5.2 Architecture, deployment, and maturity

The technical core follows Trino's coordinator-worker pattern. The coordinator parses, plans, and schedules SQL; workers retrieve and process data through source-specific connectors; catalogs hold the configurations that determine how those connectors expose each source. This makes connector behavior, pushdown, source-system capacity, and network locality part of realized performance rather than implementation detail. Starburst's current Enterprise page claims more than 50 native connectors and up to two-times faster performance than open-source Trino, while a company-commissioned ESG report cites an internal benchmark of up to 28% over standard Trino and alternatives. Neither disclosure provides the workload-level evidence needed to generalize to a buyer's estate. The Trino GitHub release page listing version 484 on 2026-10-16 does, however, provide a current developer signal for the upstream engine.[CE002, CE004, CE006, CE009, CE012]

5.3 Differentiation, trust, and technical risks

Product differentiation is the governed commercial layer around open Trino, not a proprietary storage substrate. Starburst describes row- and column-level controls, masking, encryption, audit logging, and integrations with Ranger, Vault, Okta, and LDAP. SiliconANGLE also reported AI Workflows, a natural-language AI Agent, automated tagging, an Iceberg-oriented catalog, managed table maintenance, and workload-routing updates in May 2025. These broaden the product from query federation toward governed data and AI operations, but announcement coverage does not establish production maturity or certification. Independent review evidence is directionally useful but thin: PeerSpot's comparison exposes only two Starburst reviews, SelectHub aggregates 132 reviews across two sites, and a Reddit thread raises the commercial-value question without supplying practitioner answers. Deployment complexity, resource consumption, permissions administration, and complex-query performance therefore remain diligence topics rather than proven defects across the installed base.[CE010, CE011, CE013, CE014, CE015]

5.4 Evidence gaps and underwriting implications

The retained evidence supports a credible product boundary and a live upstream project, but it does not close the underwriting questions that determine whether Starburst's premium over self-managed Trino is durable. The strongest quantified implementation and performance evidence comes from a validation commissioned by Starburst, while independent review samples either lack depth or have small, disclosed denominators. The fetched Gartner text contains only its general reviewer disclaimer and no product-specific rating detail, so it cannot be used as a current satisfaction benchmark. Diligence should obtain reproducible connector benchmarks, deployment staffing and migration records, service objectives, recovery testing, current security attestations, penetration-test results, and incident history. Absence of those records from this pool is an evidence gap, not proof that a control is ineffective or that an incident has not occurred.[CE006, CE007, CE009, CE013, CE014, CE015, CE016]

5.5 Exhibits

6.1 Buyer, user, and deployment segments

Starburst sells into organizations whose data is distributed across cloud, on-premises, and hybrid systems, with the strongest observable fit in regulated and data-intensive environments. Its buyer is typically a data-platform, analytics, or governance owner; users include analysts, data engineers, and AI teams; and the economic sponsor is motivated by avoiding data movement, improving governed access, or reducing infrastructure work. Starburst Enterprise is positioned for private-cloud, hybrid, on-premises, and air-gapped deployments, while Galaxy is the managed option. The official product surface names financial services, healthcare, insurance, and government as use cases, but the public material does not disclose segment-level revenue, account counts, or geographic mix. This supports a segment thesis, not a customer-base census.[CU001, CU002, CU003, CU004]

6.2 Adoption and named customer proof

The clearest named deployment is Citi: SiliconANGLE reports that Citi uses Starburst to unify data access across 165 countries while maintaining data sovereignty. This is meaningful production-style proof of a high-complexity, regulated use case, but it does not disclose contract value, deployment scope, renewal status, or the number of users. Commissioned Enterprise Strategy Group research adds customer-reported operating outcomes, including a newly acquired subsidiary connected and analyzed within two weeks, up to 60% less platform-management time for one customer, and 70% less infrastructure-management time for one Galaxy customer. Because those customers are anonymized and the study was commissioned by Starburst, the results are useful directional evidence rather than independently auditable cohort data.[CU005, CU006, CU007, CU008, CU009]

6.3 Retention, expansion, and concentration

Public evidence is insufficient to quantify retention or expansion. PeerSpot reports two Starburst Enterprise reviews and a 100% willingness-to-recommend figure, while SelectHub aggregates 132 reviews from two recognized sites and reports an 87% satisfaction rating; these are review signals, not renewal, GRR, NRR, or revenue retention. The available reviews also identify deployment complexity, resource consumption, difficult permission management, and usage-based cost forecasting as adoption friction. UpGuard’s October 23, 2026 vendor-risk page describes continuous monitoring, but its free extract withholds the rating and underlying findings. There is therefore no public basis in the assigned evidence for a current customer-count, churn, renewal, or top-account concentration number. These gaps matter because a regulated-enterprise sales motion can be durable while still depending on a small number of large accounts or high-touch deployments.[CU010, CU011, CU012, CU013, CU014]

6.4 Customer evidence exhibits

The exhibits separate named production evidence, anonymized outcome evidence, and broad review or segment signals. They should not be read as a funnel: the sources use different populations, units, and time bases, and none supports a conversion rate from prospects to active customers. The practical diligence priority is to obtain a customer cohort schedule showing active accounts, renewal dates, GRR/NRR, expansion revenue, implementation status, and the share of ARR attributable to the largest customers.[CU005, CU010, CU014]

6.5 Exhibits

7.1 Legal, regulatory, and governance exposure

Starburst markets deployments for financial services, healthcare, insurance, and government, including air-gapped and data-residency-sensitive environments. That positioning raises the consequence of access-control, masking, audit-log, and data-sovereignty failures even where no public enforcement event was found in the assigned evidence. Its own Enterprise page describes row- and column-level controls, encryption, third-party integrations, and audit logging, while independent material describes governance and compliance as product benefits rather than as an audit result. The investment implication is conditional: regulated-market differentiation is valuable only if controls, certifications, incident response, and customer contractual obligations can be verified in diligence.[CR001, CR002, CR003, CR004]

7.2 Operational, quality, and security risks

The platform's architecture distributes work across coordinators, workers, connectors, catalogs, and remote data systems. This creates multiple failure surfaces: query planning, connector behavior, source availability, resource sizing, permissions, upgrades, and performance under concurrency. A current job posting describes production support, migration troubleshooting, platform health, query failures, performance tuning, and incident root-cause analysis as required operating work. Independent review evidence also flags resource consumption, complex setup, slow complex queries, and decentralized permissions as user-reported or analyst-identified limitations. These signals do not prove an incident, but they support a material execution risk and make uptime, query-error, recovery-time, and security-incident metrics important kill criteria.[CR005, CR006, CR007, CR008, CR009]

7.3 Partner, platform, and financial-model dependencies

Starburst depends on the Trino open-source engine, a broad connector ecosystem, cloud and on-premises infrastructure, and customer teams capable of administering distributed data systems. PeerSpot's comparison says Starburst can require more technical knowledge to deploy, while the company's own material emphasizes more than 50 connectors and operation across multiple environments. That combination creates both differentiation and support burden: connector compatibility, upstream vulnerabilities, cloud capacity, and customer skills can affect time to value. Public evidence in this pool does not establish cash runway, burn, margin, debt, customer concentration, or the structure of the Citi investment, so financial resilience and partner concentration remain unresolved rather than assumed safe.[CR010, CR011, CR012, CR013]

7.4 Exhibits

8.1 Investment case and price discipline

Starburst has a credible enterprise data-infrastructure proposition: it commercializes Trino-based federated querying, sells both managed and self-managed offerings, and positions governance and data residency as buying criteria for regulated customers. Independent coverage reported a $250 million Series D in February 2022 that brought the company to a $3.35 billion valuation and $414 million of cumulative financing, while the same reporting described nearly three-times revenue growth in 2021. Those facts support a real scale and financing signal, not a current valuation mark. The 2025 SiliconANGLE report adds a strategic Citi investment and says Citi uses Starburst across 165 countries, but gives no investment amount or price. Public evidence therefore supports a quality-positive, price-conditional thesis rather than a buy decision. Entry discipline should require a current cap table, preferred terms, latest financing price, ARR bridge, gross margin, net retention, cash, debt, and burn. Without those inputs, a precise private-company multiple or target return would be false precision.[CV001, CV002, CV003, CV004, CV005]

8.2 Thesis, anti-thesis, and comparable evidence

The upside thesis is that Starburst can monetize a durable need to query distributed and sensitive data without centralizing it, then expand from query federation into governed data products and AI workflows. Sources describe more than 50 connectors, enterprise controls, and a product surface spanning Galaxy and Enterprise; these are differentiation signals, but most are company or ecosystem descriptions rather than audited financial proof. The anti-thesis is that open-source Trino limits switching costs, cloud data platforms and adjacent vendors can compress pricing, and the company may need sustained sales and engineering investment to turn a broad platform into durable recurring economics. A 2026 Enterprise Strategy Group validation commissioned by Starburst modeled 45% lower three-year total cost of ownership and a 414% return on investment for a hypothetical customer, but that sponsored model is not evidence of Starburst's own margins or valuation. The only directly refreshed public-market reference in the assigned pool is Snowflake, whose October 2026 market capitalization was reported at $110.76 billion; that is a scale and public-liquidity benchmark, not a like-for-like private valuation multiple. The comparable set is consequently directional and cannot justify marking Starburst above or below $3.35 billion without current Starburst revenue and margin data.[CV006, CV007, CV008, CV009, CV010]

8.3 Scenarios, exit readiness, and diligence gates

A scenario framework is more reliable than a point estimate. In the bull case, Starburst converts its distributed-data and regulated-AI positioning into sustained expansion, with retention and margins strong enough to support a premium software multiple. In the base case, it remains a valuable but execution-sensitive private platform whose fair entry price depends on verified ARR quality and dilution terms. In the bear case, cloud competition, open-source substitution, slower AI monetization, or a down round compresses value and lengthens the exit window. These are conditional investment signals, not probability-weighted forecasts, because public sources do not disclose current ARR mix, GRR, churn, customer concentration, profitability, cash, debt, or the Citi transaction structure. Exit readiness is therefore unproven: historical reporting mentioned an IPO ambition, but no current filing or liquidity timetable is evidenced here. The decision should move toward buy only after price and operating evidence close the listed gaps; otherwise track or research-more is the appropriate protection against private-market information asymmetry.[CV011, CV012, CV013, CV014, CV015, CV016]

8.4 Exhibits